Dense Passage Retrieval for Open-Domain Question Answering
This addresses the efficiency and accuracy of retrieval for open-domain QA systems, representing a significant advancement over traditional sparse methods.
The paper tackled the problem of passage retrieval for open-domain question answering by proposing a dense retriever using learned embeddings, which outperformed a strong BM25 system by 9%-19% in top-20 accuracy and achieved new state-of-the-art results on multiple benchmarks.
Open-domain question answering relies on efficient passage retrieval to select candidate contexts, where traditional sparse vector space models, such as TF-IDF or BM25, are the de facto method. In this work, we show that retrieval can be practically implemented using dense representations alone, where embeddings are learned from a small number of questions and passages by a simple dual-encoder framework. When evaluated on a wide range of open-domain QA datasets, our dense retriever outperforms a strong Lucene-BM25 system largely by 9%-19% absolute in terms of top-20 passage retrieval accuracy, and helps our end-to-end QA system establish new state-of-the-art on multiple open-domain QA benchmarks.